maximum number of columns in an object

[iliyansomlev]

Hi,

Do you know what is the greatest number of columns we can have in a chart (or table box) in QV 11? I have a straight chart with 150 columns. Is it possible to have more columns like, say 400 or 600 columns?

Thanks

[#2]

Hi,

Look here, can help:

http://community.qlik.com/message/321120#321120

By Rebeca

[Gysbert_Wassenaar]

As far as I know the theoretical limit is extremely high. The practical limit depends on the amount of ram and cpu power of your machine. Add enought columns and it'll perform so bad you'll want to take it behind the barn and shoot it. Except you won't be able to because it won't move an inch anymore.

[hic]

Gysbert Wassenaar is totally right. Also about the barn...

Bottom line is:You should normally not have more than 4 dimensions. It depends on the number of distinct values in each of the dimensions. So sometimes you can have more, sometimes you must settle with less. If you have 4 dimensions with each 200 distinct values, then QlikView must calculate a maybe complex aggregation 200x200x200x200 times, i.e. 1.6 billion times. The barn is coming closer...

Adding expressions is cheaper though. It is in principle linear, so that if you double the number of expressions, you will double the response time.

HIC

[iliyansomlev]

Thank you for your answers, they helped me a lot.

It is good not to have limits even when this refers to theoretical limits only

In my case I have about 40 dimensions and 100 expressions in a straight table object and all of them are conditional and selected from a list box. The idea was to give maximum flexibility of the end user and leave number of selected dimensions to their common sense. Maybe I should limit this number as a calculation condition. For a limited subsample of 100 or so rows there was no problem having all 140 columns in the table calculated quite fast, given that several expressions use aggr function.

[#6]

This is great information. Keep it up and I hope the original poster will summarize and post an answer and accept one as well.

[#7]

Are you referring to data cloud tables and their strength, not much of source columns as source columns? Because sources will eventually get "associated" or "related to" one another.